COMP S834. Unit 5. Databases and session management

Transcription

1 COMP S834 Unit 5 Databases and session management 120

2 Course team Developer: Designer: Coordinator: Member: Jenny Lim, Consultant Chris Baker, OUHK Dr Li Tak Sing, OUHK Dr Andrew Lui Kwok Fai, OUHK External Course Assessor Prof. Mingshu Li, Institute of Software, Chinese Academy of Sciences Production ETPU Publishing Team Copyright The Open University of Hong Kong, Revised All rights reserved. No part of this material may be reproduced in any form by any means without permission in writing from the President, The Open University of Hong Kong. Sale of this material is prohibited. The Open University of Hong Kong 30 Good Shepherd Street Ho Man Tin, Kowloon Hong Kong

3 Contents Overview 1 Introduction 3 Database concepts 4 Relational databases 6 Database management systems (DBMS) 9 Structured Query Language (SQL) 13 PHP and MySQL 20 ABC Books: our case study business 20 PHP MySQL functions 20 Open Database Connectivity (ODBC) 23 Database abstraction 27 Session management mechanisms 30 Client-side techniques 31 Server-side techniques 36 Issues in state management 43 Ensuring security 43 Managing sessions across multiple servers 43 Summary 45 Suggested answers to self-tests 46 References 50

4 Unit 5 1 Overview Databases are used in any information system which needs to store, organize and manage large amounts of data. They also play an important role in making an organization s information resources available online. Product catalogues, customer orders and user registrations are just some of the types of data that can be stored in back-end databases which are connected to the Web. In the first part of this unit, we ll briefly look at the history of databases. We also look at the most common data model used by database systems today: the relational model. We then discuss the ways in which users can access Web databases and will look at the basic SQL commands that are used for retrieving and updating information in a database. We will also look at how these commands can be incorporated within a server-side script written in a language such as PHP. The second half of the unit deals with session management. In general, a session begins when a user visits the first page on the site, and ends when he/she clicks onto a new site or remains inactive for a long period of time. There are many situations when a Web server needs to keep track of what has been going on within a user session. For example, search engine sites must know the page you are currently viewing so that they can retrieve the previous or next page of results correctly. Shopping sites need to match online shoppers with their virtual shopping baskets through all the pages that they visit. However, HTTP is a stateless protocol, which means that a Web server processes each request for a new webpage as a discrete unit, without retaining any knowledge of previous pages requested within the same user session. This design approach leads to better performance and scalability. However, many Web applications depend on the ability to track individual requests from distinct users and store information about the current state of a session for each client. We will look at a number of techniques which are used to store session information for HTTP clients. We will also discuss the pros and cons of each of these methods. These techniques are all implemented within the application level due to the lack of support for it in the underlying protocol. Using database technology and the session management techniques you ve learned, you will implement a shopping cart application for ABC Books. This will allow shoppers to browse and select books from a product catalogue database running behind their website. Session management techniques will be used to identify each user and to keep track of the items in their shopping cart. Once again, we will use PHP and MySQL to implement this application.

5 2 COMP S834 Web Server Technology In short, this unit: describes how databases can be integrated into Web applications; discusses the necessary technologies for implementing Web-database connectivity; examines how session management can be performed within a stateless protocol like HTTP; helps you to construct a Web application which interfaces with databases and maintains state over a series of HTTP requests; and explains relevant issues when planning and adopting a session management strategy for a website. This unit should take you about five weeks or hours to complete. Please plan your time carefully and remember to contact your tutor if you have problems.

6 Unit 5 3 Introduction Static HTML pages are adequate for publishing small amounts of information that does not change frequently. By contrast, dynamic HTML pages are required for publishing data which changes very often or when your users need to enter and retrieve live data from a database using a form. Most e-commerce sites nowadays make use of back-end databases to store their content. Server-side programs are then used to assemble the data from the database in many different ways in response to specific user requests. Relational database management systems are the most common technology used for storing, organizing and managing Web databases. Server-side programs can issue calls against the database using standard technologies such as Open Database Connectivity (ODBC) or through built-in functions provided by the scripting language. Requests for retrieving, updating, or adding records to the database are written using another standard known as SQL (Structured Query Language). The Web server interfaces with server-side programs and application servers which, in turn, issue the requests to back-end databases. Web server Web browser Completed form HTML result HTTP Server-side script Send query to DB Build HTML with results Query Results Database Figure 5.1 Web-database connection To implement shopping carts, user preferences, or detailed visitor tracking, the server must be able to recognize a user throughout an entire session, from one page to the next. This is the essence of session management. As mentioned in the Overview, session management is not so straightforward under HTTP and must be implemented within the Web application itself. We will discuss both client-side and server-side techniques for managing user sessions. Client-side techniques, such as cookies, involve storing the session data on the client machine. Server-side techniques involve storing the session state within a file or database on the Web server itself. That s right, databases are used not just for storing and organizing website content, they also play a very important role in session management! Implementing the session management techniques discussed in this unit will help you deliver a richer, more interactive Web environment to your users.

7 4 COMP S834 Web Server Technology Database concepts We ll start this section with a brief history of databases. We will also look at the most common data model used by database systems today: the relational model. We ll discuss the core concepts of the relational model and show you how data is stored and organized based on this model. Towards the end of the section, you will get a chance to perform data administration and manipulation tasks using MySQL, a very popular open-source database management system which uses the relational data model. Flat files The business world started adopting computers for their data processing needs after World War II. Using the same approach as scientists and engineers, businesses originally stored data in flat files where data was organized in tabular format. Rows in the table were called records and columns were called fields. (In Figure 5.2, records give information about individual students while fields deal with categories such as students names, subjects or grades.) Marvin Kwok 2003 Semester 1 Chemistry 101 C Alice Tam 2003 Semester 2 Physics 102 B- Betty Cheung 2003 Semester 2 Math 150 B Figure 5.2 A subset of a file containing student results In order to improve the performance of flat files, indexes were built on selected fields which are commonly used to search the file. These index files allowed applications to access records directly rather than searching through the entire file from the very beginning. Flat files were not well-suited for the storage and management of significant volumes of data. A flat-file structure only contains a single table and is not practical for most business applications, which must store information about multiple entities (e.g. Customer, Order, Invoice, Supplier) that may have relationships with each other. In addition, early data-processing systems tended to grow independently without an overall plan to guide their development. Departments such as accounting, finance, manufacturing, human resources, sales and marketing would develop their own programs and data files in isolation from other functional areas. This approach led to multiple master files which were created, maintained and operated by separate divisions or departments. The next figure illustrates how data is organized within the traditional flat-file environment.

8 Unit 5 5 Prog A Prog B Prog C Prog A Prog B Prog C Order-filing-system Accounting-system Customer- Master- File Inventory- Master- File Order- Master- File Customer- Master- File Inventory- Pricing- File Figure 5.3 Traditional flat-file approach Several problems result from a flat-file approach: 1 Data redundancy and confusion The same employee, customer or product record may appear in several files belonging to different departments. This makes it difficult to keep the same information consistent throughout the organization. 2 Program-data dependence There is a tight relationship or coupling between data stored in flat files and the specific programs required to update and maintain those files. Any change in the file structure, such as adding a new field or increasing the length of an existing field, requires a change in all programs that access the data. 3 Lack of flexibility End-users did not have any tools to access flat files directly. The specialized skills of a programmer were needed in order to write a program which could extract the data and generate the necessary reports. This made it difficult to deliver ad-hoc reports or respond to unanticipated information requirements. Database technology can solve many of the problems created by the traditional flat-file approach. A database is actually a collection of data organized to serve many applications efficiently by centralizing the data and minimizing redundant data. Databases may contain one or more tables. In effect, a single database can serve multiple applications simultaneously. The next figure illustrates the shared database approach.

9 6 COMP S834 Web Server Technology Management Manufacturing Sales and Marketing Corporate database Accounting Human Resources Customer Service Figure 5.4 Shared database environment Databases can be classified by the model or data structures associated with them. Some models have been in wide use for some time, such as the hierarchical and network models. However, the dominant model today is the relational model, used together with the SQL query language. This is the model that we will adopt for the ABCBooks database. Relational databases Relational databases are based on the relational model, which was invented by an IBM researcher named Ted Codd in The relational model is based on relational calculus and algebra, and was meant to allow non-technical users to store and retrieve large amounts of information. Ted Codd envisioned a system where the user would be able to access information with English-like commands, and where information would be stored in tables. Relational databases actually represent data as a collection of twodimensional tables. A table can be visualized as a group of records which, in turn, are made up of fields. Tables are also referred to as files, since each table can be viewed as equivalent to a flat file. However, databases may contain one or more tables, while flat files only hold one table at a time. The next figure illustrates this hierarchy.

10 Unit 5 7 Student database Course enrollment file Payment file Personal history file Course enrollment file Student John Woo Betty Chan Iris Cheung Course IS 100 IS 100 IS 100 Term 2003 Sem Sem Sem3 Grade B+ A B- Course enrollment record Name John Woo Course IS 100 Term 2003 Sem2 Grade B+ Figure 5.5 Data hierarchy in a relational database Each table stores data about a single entity, which could be a person, place, object, event or concept of interest to the business. Books, customers and orders are just some of the entities that will be stored in the ABCBooks database. The rows within a table correspond to the records that contain information about a single instance of an entity. For example, there will be one record within the ABCBooks Book table for A Dream of Red Mansions by Tsao Hsueh-Chin, which is one particular instance of a book title. The fields within a table contain the characteristics or attributes recorded for each entity. Databases generally support three types of fields: numeric types, date and time types, and string or character types. There is a special field known as the primary key, which serves to identify each record uniquely within a table. book ID, title, description, listprice and monthpublished and yearpublished are just some of the fields that may be stored for each book, with the book ID serving as the primary key. The next figure shows an example of the tables and fields for our initial database design for ABC Books. The primary keys are underlined for each table.

11 8 COMP S834 Web Server Technology Customer (customerid, lastname, firstname, middleinit, address1, address2, district, region, phone, , password, join_date, last_login_date) Order (orderno, customerid, dateplaced, dateshipped, totalamount, Status) Order Detail (orderno, orderlineno, bookid, unitprice, quantity, discount) Book (bookid, title, auth_firstname, auth_lastname, auth_middlename, monthpublished, yearpublished, numpages, listprice, description) Author (Author_FirstName, Author_LastName, Author_MiddleName, Biography) Figure 5.6 Sample database design Relational databases can also record the relationships or linkages between tables through common fields which occur in the related tables. In the example above, there is a relationship between the Order table and the Customer table through the field customerid. We can therefore retrieve all the orders for a particular customer by looking for all Order records that contain a specific customerid. We can also look up the customer details for a particular order by using the same ID. customerid is considered a foreign key within the Order table, because it is a primary key in another table. Foreign keys are used to look up information in another table to avoid keeping the same data in more than one place. Here are the three basic operations used to extract data from a relational database: 1 Select creates a subset of records from a file which meet a certain criteria. For example, we may want to select all books written by Mark Twain from the Product table. 2 Join combines fields from more than one table. For example, we may want to join the Product table with the Order Line table so that we can see all the products within an order, along with their descriptions, number of pages and date of publication. 3 Project permits the creation of new tables or views that contain only the information required from the underlying table or tables. For example, some programs may not need all the fields in the Customer table, so a view of this table containing only the fields required may be created and made accessible to these programs. The actual commands for these operations are expressed in Structured Query Language, which you ve used briefly in Activity 4.11 (about fulltext searching with MySQL). The users and application programs who issue these commands do not need to know how these operations will

12 Unit 5 9 actually be carried out. They just need to issue the right instructions and the database will take care of servicing the request. Because of this, relational databases are much more flexibile in accomodating ad-hoc queries, or queries which combine information from different tables and sources. The looser coupling between programs and data also allows the addition of new tables, columns, and records without the need to rewrite the programs that access the database. This has made relational databases very popular with businesses. Self-test Briefly describe the problems caused by using the flat-file structure for storing and organizing data. 2 How is data organized using the relational data model? 3 What is the difference between a primary key and a foreign key? 4 Referring to Figure 5.6, identify two other foreign keys aside from CustomerID. What relationships are established by these keys? Database management systems (DBMS) A database management system is simply the software used to create, access and manipulate databases. The DBMS acts as an interface between application programs and the physical data files. It relieves the programmer or end user from the task of understanding where and how the data is actually stored by separating the logical and physical views of the data. A database management system has three major components: 1 Data Definition Language (DDL) language used to define the content and structure of the database, such as creating and dropping databases and tables. 2 Data Manipulation Language (DML) language used to extract, insert, update and delete data. 3 Data dictionary inventory of all the data elements or fields within the database; it contains metadata or data about the database itself. This allows programs and users to query the server and find out what data structures (e.g. databases, tables, indexes, fields) are stored within. Many popular database management systems operate using the clientserver model, with a database server responding to client requests over a

13 10 COMP S834 Web Server Technology network connection. However, it is also possible for the client and database server to be on the same machine. User 1 User 2 Application 1 Application 2 Data definition Data manipulation Data administration DBMS engine Database Data dictionary Figure 5.7 DBMS serves as an interface between applications, users and the physical data files We will be using MySQL to implement the back-end database for user registration, catalogue search and online ordering on the ABC Books website. The next reading gives you a quick overview of the programs that you will use to administer and manipulate MySQL databases. These programs make up the MySQL DBMS, and are classified as server, client or utility programs. You will use some of these programs in the following activities. Reading 5.1 Overview of MySQL programs, Section 4.1 in MySQL Manual, These programs can be invoked at the command line by entering the program name, followed by any options or other arguments needed to instruct the program. Please note that mysqld is the name of the MySQL database server, and this program must be running before clients can connect to it. In the next activity, we will explore some of the administrative functions that can be performed with MySQL. Activity 5.1 We will now perform some common database administration tasks such as starting and stopping the database server, creating a new database and adding a new user account. We will use three MySQL client programs in this activity mysql, mysqladmin, and mysqlshow. They were all briefly described in Reading 5.1. While you re working through the examples below, you can

14 Unit 5 11 also refer to this URL as a more detailed reference for these and many other useful programs: MySQL client and utility programs, Just a quick reminder before you start typing commands into the interactive MySQL client programs. You re free to type your SQL commands in upper or lower case, but if you re using MySQL on a UNIX-based system, be aware that database and table names are casesensitive. This is because they correspond to directories and files in the MySQL data directory. So if you defined a table as Order_Detail and then try to query it using the name order_detail, MySQL will tell you that the table doesn t exist. Part 1 Starting and stopping a server There are also a number of ways to stop the server, but we will use the script /etc/rc.d/init.d/mysqld to do this. 1 Check whether MySQL is already running: shell> mysqladmin status 2 If MySQL is already running, you can shut it down: shell> /etc/rc.d/init.d/mysqld stop 3 You can restart MySQL using the same script. shell> /etc/rc.d/init.d/mysqld start 4 Now use mysqladmin to check the status of the MySQL server after restarting it: shell> mysqladmin status Part 2 Creating a new database 1 We will now create a new database called ABCBooks using the mysqladmin client program: shell> mysqladmin create ABCBooks 2 Next, verify that the database has been created using the mysqlshow command. This command will give you a listing of all the existing databases on the server. You should see ABCBooks among the databases listed: shell> mysqlshow

15 12 COMP S834 Web Server Technology Part 3 Adding a new user account MySQL is initially installed with default accounts for the root user. These accounts have blank passwords, so anyone can log in to the database as long as they know the account name. For security reasons, however, you do not want your Web server scripts to connect to the database server as root. We will create a new user account to be used by our scripts for accessing and manipulating the ABCBooks database. 1 You will use the mysql client to connect to the MySQL server as root. The mysql client is an interactive program that allows you to connect to a MySQL server, run queries, and view the results: shell>mysql u root Note that if you have already assigned a password to root (which is a good idea), you should include the password using the -p option, with no space between the option switch and the password (i.e. -pmypassword). 2 Once you ve logged in as root, you can select which database you re going to work on from the several databases that the server currently maintains. Since we re going to create a new database user, we need to use the system database called mysql where user accounts are stored: mysql shell> use mysql; 3 User accounts are stored in a table called user within the MySQL database. We will insert a record for our new user into this table. After adding the record, you must tell the server to reload the user account information with FLUSH PRIVILEGES. Without issuing the FLUSH PRIVILEGES command, the server will not recognize the newly-added user until after the server has been restarted. Lastly, we will use the PASSWORD function to encrypt the password before storing it in the user table: mysql shell> insert into user (Host, User, Password) values ( localhost, webuser, PASSWORD( comps834 )); mysql shell> flush privileges; mysql shell> grant select, insert, update, delete, index, alter, create, drop on ABCBooks.* TO localhost identified by comps834 ; We have just created a new user called webuser who is allowed to perform data manipulation (e.g. insert, update, select and delete) and administration (e.g. create and drop) tasks on the ABCBooks database. 4 You can verify that the user has been added correctly by logging into the MySQL client once more using this new account. Exit from the MySQL client command shell with:

16 Unit 5 13 mysql shell> \q Then from the operating system command line, you can log in as the new user. Please note that there must be no space between the -p option switch and the password. You can also specify the database that you want to use (e.g. ABCBooks) as part of the options in the command line up front: shell>mysql u webuser pcomps834 ABCBooks Our ABCBooks database is empty for the moment, but it won t stay that way for long! Now do the following self-test to revise the database management concepts you ve learned in this section. Self-test 5.2 For questions 2 and 3 below, you may want to consult this reference: 1 Describe the steps involved in adding a new database user account to MySQL. 2 Describe four values that are displayed when you run the mysqladmin status command. 3 Write the mysqlshow command that will display information about the Author table in the ABCBooks database. Structured Query Language (SQL) SQL (pronounced ess-que-el ) is a standard, declarative language used to communicate with a relational database. It is declarative because it describes what must be done rather than how it must be done. Application programmers who use SQL statements do not have to worry about how the database will carry out their request. They just have to issue commands using SQL, send the commands to the database server, and the DBMS will carry them out. SQL is used to implement the DDL (Data Definition Language) and DML (Data Manipulation Language) of many popular relational DBMSs such as MySQL, Oracle, Sybase, Microsoft SQL Server, Microsoft Access, Postgress, etc. Be aware that although most database systems use SQL, they may also have their own additional SQL extensions that are available only on their system. This means that different DBMSs may not provide the complete

17 14 COMP S834 Web Server Technology set of SQL functions or may include functions which are not part of the standard. In this unit, we will discuss the standard SQL commands such as Select, Insert, Update, Delete, Create, and Drop. This small set of commands is adequate for accomplishing most of the actions that we will perform on the ABCBooks database. The next reading goes through each of these SQL commands in detail. You can also try submitting your own SQL statements in an online form at the end of each SQL command discussed. Reading 5.2 You may skim through these readings first, just to get a fair idea of what s involved before you start entering SQL commands interactively in the MySQL server on your local machine: 1 Selecting data, 2 Creating tables, 3 Inserting into a table, 4 Updating records, 5 Deleting records, 6 Drop a table, In the next activity, we will issue SQL commands to create and manipulate the tables within the ABCBooks database. The commands will be entered via the MySQL client program, which is an interactive program that allows you to connect to a MySQL server, run queries, and view the results. Activity 5.2 The files needed for this activity can be downloaded from the course website. If you need help while completing the tasks, you can refer to the following portions of the MySQL online documentation: SQL statement syntax, Column types (or types of data supported by MySQL), 1 Use the mysql client to log in to the MySQL server as webuser. Specify that you will be working on the ABCBooks database. (Refer to Activity 5.1 for more information on how to do this.)

18 Unit Now type the following SQL statement into the mysql client command shell. This will create a table called Book, which stores information about the titles sold by ABC Books. The semicolon on the last line denotes the end of the statement. Pressing the Enter key after typing the semicolon will cause the shell to execute the statement: create table Book ( bookid int NOT NULL AUTO_INCREMENT PRIMARY KEY, title varchar(200) NOT NULL, author_lastname varchar(150) NOT NULL, author_firstname varchar(150) NOT NULL, author_middlename varchar(150), imageurl varchar(100), price numeric(8,2), numpages numeric(5), pub_month numeric(2), pub_year numeric(4), description text, fulltext (title, author_lastname, author_firstname, author_middlename, description) ); This statement creates a table named Book with 11 fields. bookid is used as the primary key, which means that an index will be created for this field and no two rows of the table will have the same bookid. A unique number will automatically be generated by MySQL and stored in the bookid field for every new record inserted into the table. title, auth_fname, auth_lname, auth_mname and imageurl are strings of variable lengths. The longest strings that can be stored within these fields are indicated by the numbers within the parentheses. listprice can contain a value with two decimal places. numpages, monthpublished and yearpublished can only contain integers with the specified maximum number of digits, while description can store a very long block of text. 3 Aside from the book description, ABC Books also wants to store short biographies of their authors in the database. We will create a new table called Author so that we can store author-related information separately from the books they ve written. It is more efficient to store one copy of the author s record in a separate table called Author, and then link the Book and Author table through the foreign key fields auth_firstname, auth_lastname and auth_middlename. Let s create the Author table next:

19 16 COMP S834 Web Server Technology create table Author ( auth_lastname varchar(150) NOT NULL, auth_firstname varchar(150) NOT NULL, auth_middlename varchar(150) NOT NULL, auth_bio mediumtext, primary key (auth_lastname, auth_firstname, auth_middlename) ); 4 Next, let s verify that the two tables were created with this command: show tables; We can also check the fields that have been created for these tables. The two commands shown below will be executed separately, one after the other (note the semicolon at the end of each statement): describe Book; describe Author; 5 We have a few more tables to set up, such as Order, Order_Detail and Customer. The CREATE statements for these tables are already in a text file called create_abc.sql, and instead of typing in these statements all over again, we can just tell MySQL to read and execute the contents of the input file for us: source filename; where filename contains the name of our input file, create_abc.sql. 6 Now that all our tables have been created, let s start loading some records into them. The Customer table contains information about users who have registered on the site. It also makes use of an autogenerated ID field to uniquely identify each user, same as the Book table. There are comments in the SQL statement explaining the fields in the Customer table, denoted by two forward slashes (//). These comments are for your information only, and you do not have to enter them into the shell: insert into Customer (firstname, lastname, address1, address2, district, region, phone, , password, join_date) values( Jenny, //first name Lim, //last name N, //middle init Flat D, 22/F Healthy Gardens, address1 305 King s Road, //address2 North Point, //district Hong Kong, //region , //phone jennylim@fre .com, // password( secret ), //password now() //current date );

20 Unit 5 17 Once again, we will use the MySQL password function to encrypt the password in the database. We also keep track of the date when the customer registered on the site with the MySQL now function. 7 Now, you can retrieve the record you just added by using the SELECT command: select * from User where = jennylim@fre .com ; The asterisk(*) denotes that all fields within the table should be retrieved. You can also specify the field names that must be returned, separated by commas. 8 The file load_abc.sql contains more INSERT statements for loading entries into the Book, Author and Customer table. You can run this script file using the source command: source load_abc.sql MySQL would treat the contents of load_abc.sql as commands to be executed. 9 Now, you can check the contents of the two tables by typing the following commands: select * from Book; select * from Author; select * from Customer; 10 Another interesting thing to try is the JOIN operation, which allows you to query information from more than one table. Let s create a SELECT statement which displays the author s biography along with the rest of the information about a book: select bookid, title, auth_firstname, auth_lastname, auth_middlename, listprice, monthpublished, yearpublished, numpages, description, auth_bio from Book, Author where Book.author_lastName = Author.auth_lastName and Book.auth_firstName = Author.auth_firstName and Book.auth_middleName = Author.auth_middleName; 11 Next, let s try the UPDATE and DELETE commands. You can verify that the commands were executed correctly by using SELECT to query the tables before and after issuing the UPDATE and DELETE commands: UPDATE Book SET listprice = where bookid = 5; UPDATE Book SET listprice = listprice *.75 WHERE bookid = 10; DELETE FROM Author WHERE auth_firstname = Eileen and auth_lastname = Chang ;

21 18 COMP S834 Web Server Technology 12 Now, let s try creating indexes on some fields other than the primary key. Indexes can speed up searching on fields which are commonly used as search criteria, such as the book title and the customer s first and last name: CREATE INDEX idx_title ON Book (title); where idx_title is the name of the index, Book is the table and title is the field being indexed. 13 Here s another command which indexes the Book table on the Description field. This field is used for full-text searching, so we have an additional keyword FULLTEXT in the command: CREATE FULLTEXT INDEX idx_ description ON Book (description); 14 Another command that can be used to create an index is the ALTER TABLE command: ALTER TABLE Book ADD INDEX idx_author (auth_lastname, auth_firstname, auth_middlename); 15 The file index_abc.sql contains more statements for creating indexes for the tables in the ABCBooks database. You can run this script file using the source command: source index_abc.sql 16 You can exit the MySQL command shell by typing quit or \q. There are many other useful SQL commands we haven t tried yet. However, the commands you ve learned in this activity should be enough for the majority of the database operations you ll need to perform. Please refer to the SQL statement syntax reference provided at the start of this activity for a complete list of SQL commands available in MySQL. As a summary, here are the SQL statements discussed in this activity: 1 Data definition (DDL) CREATE TABLE; and ALTER TABLE (changing field type and length, adding an index). 2 Data manipulation (DML) INSERT; UPDATE; DELETE; SELECT (single table); and

22 Unit 5 19 SELECT (join). You ve gone through a lot of database operations in this section! You ve also built some of the data structures and loaded all the necessary data into the ABCBooks database. In the next section, you will learn how to issue requests to the database from within an application program. Please complete the following self-test to assess your understanding of SQL. Self-test What will happen if you issue the DELETE statement without specifying any criteria? Example: DELETE FROM AUTHORS; 2 Write an SQL statement that will create an index on the field of the Customer table. 3 Write an SQL statement which retrieves all records from the Book table where the title contains the string Earth. 4 Identify two column types that may be used when defining date/time fields in a table. You can use this URL as a reference:

23 20 COMP S834 Web Server Technology PHP and MySQL Previously, you were issuing commands directly to the database server from a command-line client. In this section, we will incorporate the SQL commands within a PHP program which acts as a client to the database server. The PHP program runs as a server-side script on our Web server. It will process inputs from our Web users and generate dynamic pages from the database in response to their requests. We will start the section by discussing the native PHP functions for accessing MySQL databases. Later on, we will also talk about the Open Database Connectivity (ODBC) programming interface and how this can allow the same program to work with many different databases (not just MySQL). We will end the section with a brief introduction to a programming approach which uses a database abstraction layer to centralize all the database access code within your server-side scripts. ABC Books: our case study business Here s a breakdown of the functions which we will implement for ABC Books. 1 Searching for a book using a variety of methods: by title entire title or part of the title; by author various combinations of first name, last name and middle name (if any); and by keywords full-text search on book titles, book descriptions, author names and author descriptions. 2 Maintaining a shopping cart: Users can add a book to their shopping cart by clicking on an Add to Cart button. The quantity for each book is set to 1 by default but can be updated. Setting the quantity to 0 will remove the book from the cart. Once users have selected all the books they wish to purchase, they can proceed to Order Checkout. At this point, they will have to log in with their address and password before they can continue processing the order. The User Login/Registration and Order Checkout functions will be covered later when we talk about session management. PHP MySQL functions PHP comes with an extensive set of built-in functions that allow it to interface with a wide range of DBMSs, including functions for MySQL,

24 Unit 5 21 msql, PostgreSQL, Oracle, Sybase and Microsoft SQL Server. The next reading describes the most commonly used MySQL functions in PHP. Reading 5.3 Build your own database driven website using PHP & MySQL Part 4: publishing MySQL data on the Web, Note: You only need to go through pages 2 to 5. Here are the relevant functions from the previous reading, along with sample code fragments: 1 mysql_connect establishes a connection to a database. It returns the database connection handler if the connection is successful. If you want to reuse the connection even after the script has ended, you should use mysql_pconnect instead (p stands for persistent). The following sample code can be used to connect to the ABCBooks database with our webuser account. If the connection is unsuccessful, the die function will display the error message and terminate the program afterwards: $conn = mysql_connect( localhost, webuser, comps834 ) or die( Could not connect to MySQL ); Reading 5.3 also provides another way to catch errors as a result of MySQL calls. The author puts the at sign (@) in front of the function name. Then he uses an IF statement to check whether the function returns a true or false value before displaying the error statement. This approach is fine, too. 2 mysql_select_db selects the database to be used. It accepts two parameters the name of the database and the database connection handler: $conn = mysql_connect( localhost, webuser, comps834 ) or die( Could not connect to MySQL ); $selected_db = mysql_select_db( ABCBooks, $conn) or die( Could not select database ); 3 mysql_query sends an SQL query string to the database. It will return a result handler if the request is successfully executed: $conn = mysql_connect( localhost, webuser, comps834 ) or die( Could not connect:. mysql_error()); $selected_db = mysql_select_db( ABCBooks, $conn) or die( Could not select:. mysql_error());

25 22 COMP S834 Web Server Technology $result = mysql_query( SELECT * from Book ); 4 mysql_fetch_object fetches the next row in the result set as an object. The object s properties correspond to the record s fields. This function is usually used in a loop: $conn = mysql_connect( localhost, webuser, comps834 ) or die( Could not connect:. mysql_error()); $selected_db = mysql_select_db( ABCBooks, $conn) or die( Could not select database. mysql_error()); $result = mysql_query( SELECT * from Book ); while ($row = mysql_fetch_object($result)) { echo ( ID. $row->bookid., Title:. $row- >title., Date Published:. $row->monthpublished. /. $row->yearpublished. <br> ); } 5 mysql_close closes non-persistent connections to the MySQL server and returns true or false, depending on its success: $conn = mysql_connect( localhost, webuser, comps834 ) or die( Could not connect:. mysql_error()); mysql_close($conn); In the next activity, you will install these two scripts on your local Web server: search_catalog.php and maintain_cart.php. Activity 5.3 Before you do this activity, please make sure that you have installed the package php-mysql. This can be checked by the following command: rpm -q php-mysql If you have not installed this package, run the following command as root to install it: yum install php-mysql The command will download the package from the internet. So please make sure your computer is connected to the internet.

26 Unit 5 23 Database programming in PHP is easy and convenient given the wide variety of built-in database functions that are available. For MySQL, just use the MySQL functions, for Oracle, use the Oracle functions, and so on. Open Database Connectivity (ODBC) The functions discussed in the previous section provide a highly efficient way to access MySQL databases from your PHP scripts. However, they come with one major limitation. Code written using PHP s native functions for MySQL will no longer work when you switch to another database product. Consider the Search and Shopping Cart functions written for ABC Books in the Activity 5.3. If they were to switch to a PostgreSQL database in the future, the mysql_connect and mysql_query function calls would have to be rewritten using the PHP functions that are specific to PostgreSQL, i.e. pg_connect and pg_query, respectively. ODBC makes it possible for applications to access databases in a standard and uniform way, regardless of the DBMS or operating platform used by the application. It achieves this by inserting a middle layer, called a database driver, between an application and the DBMS. Application program ODBC API ODBC database driver Database Figure 5.8 ODBC serves as a middle layer between an application and the DBMS In the figure above, the application program acts as a client to the database server. An ODBC Data Source Name (DSN) (as shown in figure 5.9) containing the connection parameters for a particular data source must be defined on the client. The configuration file odbc.ini contains information about DSNs available to all users, and this is where you will add an entry for each new data source.

27 24 COMP S834 Web Server Technology [ABCBooks] Description = ABCBooks Driver = MySQL Trace = No TraceFile = Server = localhost User = webuser Password = comps834 Port = 3306 Database = ABCBooks Figure 5.9 ODBC data source for the ABCBooks MySQL database (odbc.ini) The connection parameters for a DSN are then passed to the appropriate ODBC driver, which communicates with the database server itself. Because the ODBC database drivers are loaded at run time, a user only has to install a new driver to access a new DBMS; it is not necessary to recompile or re-link the application. This approach allows a single application to access different DBMSs with the same source code. You will also need to edit odbcinst.ini after installing a new database driver. This configuration file contains information about ODBC drivers available to all users. [MySQL] Description = ODBC for MySQL Driver = /usr/lib/libmyodbc3.so Setup = /usr/lib/libodbcmys.so FileUsage = 1 Figure 5.10 MySQL ODBC database driver (odbcinst.ini) In general, ODBC simplifies connectivity to a database because applications only need to support a single, common API rather than different APIs from various database vendors. However, this approach can increase processing time due to the intermediate step needed to translate between the native database calls and the ODBC API. If performance and speed are key issues for an application, then developers may have to resort to native database APIs at the expense of portability. The next reading describes the iodbc library package on your Linux machine. iodbc is an open source, platform-independent implementation of both the ODBC and X/Open SQL specifications. The reading also comes with sample scripts for issuing ODBC calls from PHP. If you have installed a full Fedora system, then should have unixodbc installed. You can use this instead of iodbc.

28 Unit 5 25 Reading 5.4 Using PHP with ODBC A User Guide, Here are the relevant functions from the previous reading, along with sample code fragments: 1 odbc_connect establishes a connection to an ODBC dataset name (DSN). If you want to reuse the connection even after the script has ended, you should use odbc_pconnect instead (p stands for persistent): $conn = odbc_connect( ABCBooks, webuser, comps834 ) or die(odbc_error()); odbc_error is an ODBC function which returns the error state and error message. 2 odbc_prepare parses and compiles the statement, and then the odbc_execute function (see below) is used to run it later on. This is recommended for statements which will be executed several times (e.g. within a loop), as it can boost performance: $conn = odbc_connect( localhost, webuser, comps834 ) or die( Could not connect:. odbc_error()); $sql = select * from Book where listprice > ; $result = odbc_prepare($conn, $sql) or die("couldn't prepare statement:". odbc_error()); 3 odbc_exec and odbc_do these two functions mean the same thing they will execute an SQL statement on the specified database connection. odbc_exec and odbc_do can also be used to parse, compile and then immediately execute an SQL statement in one step. This approach is fine for SQL statements which will only be run once. Statements which will be run multiple times should be prepared beforehand via odbc_prepare so that they do not have to be parsed and compiled repeatedly for each execution. Example 1 odbc_execute with odbc_prepare $conn = odbc_connect( localhost, webuser, comps834 ) or die( Could not connect:. odbc_error()); $sql = select * from Book where listprice > ; $result = odbc_prepare($conn, $sql) or die ( Could not prepare:. odbc_error());

29 26 COMP S834 Web Server Technology odbc_exec ($result) or die("couldn't execute SQL: ". odbc_error()); Example 2 odbc_exec without a previous odbc_prepare $conn = odbc_connect( localhost, webuser, comps834 ) or die( Could not connect:. odbc_error()); $sql = select * from Book where listprice > ; $result = odbc_exec($conn, $sql) or die ( Couldn t execute SQL:. odbc_error()); 4 odbc_fetch_row fetches the next row from a result set that was returned by odbc_do or odbc_exec. It returns false when there are no more rows left. The resulting fetched row can be accessed via the odbc_result function: $conn = odbc_connect( localhost, webuser, comps834 ) or die(odbc_error()); $sql = select * from Book where listprice > ; $result = odbc_exec($conn, $sql) or die ( Couldn t execute SQL:. odbc_error()); } while (odbc_fetch_row($result)){ $field1 = odbc_result($result, title ); $field2 = odbc_result($result, auth_firstname ); $field3 = odbc_result($result, auth_lastname ); $field4 = odbc_result($result, auth_middlename ); $field5 = odbc_result($result, listprice ); echo Title: $field1, Author: $field2 $field4 $field3, Price: $field3\n ; 5 odbc_close takes a specific connection ID and closes this ODBC connection. You are not required to include this function since any open database connections are automatically closed when your PHP program finishes execution. In the next activity, we will rewrite the two scripts from Activity 5.3 search_catalog.php and maintain_cart.php using ODBC function calls instead of MySQL-specific function calls. Activity 5.4 Before you do this activity, please make sure that you have installed the package php-odbc. This can be checked by the following command:

30 Unit 5 27 rpm -q php-odbc If you have not installed this package, run the following command as root to install it: yum install php-odbc The command will download the package from the internet. So please make sure your computer is connected to the internet. The contents of /etc/odbcinst.ini should look like this. [MySQL] Description = ODBC for MySQL Driver = /usr/lib/libmyodbc3.so Setup = /usr/lib/libodbcmys.so FileUsage = 1 The contents of /etc/odbc.ini should look like this. [ABCBooks] Description = MySQL test database Trace = Off TraceFile = stderr Driver = MySQL SERVER = localhost USER = webuser PASSWORD = comps834 PORT = 3306 DATABASE = ABCBooks We ve discussed how databases can be accessed from PHP by using native database functions and the ODBC Standard API. With ODBC, the same program will work regardless of the back-end database used. However, ODBC could result in slower and less efficient performance due to the middle layer that is inserted between the database and the application program. In the next section, we will briefly introduce another programming approach that can be used for database access within PHP. Database abstraction The main reason for using ODBC is to avoid having to rewrite and retest your programs if you should ever switch to a new database. Another way to avoid this unpleasant situation is to create a database abstraction layer. Database abstraction will create a centralized body of code which

31 28 COMP S834 Web Server Technology handles all interaction with the database. All SQL statements are contained in this centralized body, so any changes only need to be made to this small set of code rather than all over the place. Ideally, a programmer who uses a database abstraction layer does not need to know the details of the underlying database. Even if you do not anticipate switching to another database in the future, a database abstraction layer is still a good idea. Centralizing database access results in cleaner and easier-to-manage code. One way to abstract the database is to wrap each API-specific function in a more generic function called numrows. In the example below, the mysql_num_rows function is wrapped in another function called numrows. PHP application programs would then call the numrows function instead of using the API-specific function directly. If we were to switch to PostgreSQL in the future, we would only have to port the function call from mysql_num_rows to pg_numrows in one place. function numrows($result) { //Return the number of rows in the result set. return (@pg_numrows($result)); } function numrows($result) { //Return the number of rows in the result set. return(@mysql_num_rows($result)); } Aside from writing your own database abstraction layer, there are several that can be downloaded from the Web for free. One of them is PEAR DB, which consists of a set of classes that contains our database functions. It also comes with an object-oriented style query interface. Read the following optional reading if you wish to know more. Reading 5.5 (Optional) Kent, A, Best practices: data abstraction, This concludes our section on accessing MySQL databases from PHP. Do the following self-test to check your understanding of this topic. Self-test Refer to the MySQL documentation for details on the mysql_fetch_assoc function. This can be found under the